/// NOTE: This ID is used to create part names which are then persisted in ZK and as directory names on the file system.
/// So if you want to change this method, be sure to guarantee compatibility with existing table data.
String MergeTreePartition::getID(const Block & partition_key_sample) const
{
if (value.size() != partition_key_sample.columns())
throw Exception("Invalid partition key size: " + toString(value.size()), ErrorCodes::LOGICAL_ERROR);
if (value.empty())
return "all"; /// It is tempting to use an empty string here. But that would break directory structure in ZK.
/// In case all partition fields are represented by integral types, try to produce a human-readable ID.
/// Otherwise use a hex-encoded hash.
bool are_all_integral = true;
for (const Field & field : value)
{
if (field.getType() != Field::Types::UInt64 && field.getType() != Field::Types::Int64)
{
are_all_integral = false;
break;
}
}
String result;
if (are_all_integral)
{
FieldVisitorToString to_string_visitor;
for (size_t i = 0; i < value.size(); ++i)
{
if (i > 0)
result += '-';
if (typeid_cast<const DataTypeDate *>(partition_key_sample.getByPosition(i).type.get()))
result += toString(DateLUT::instance().toNumYYYYMMDD(DayNum(value[i].safeGet<UInt64>())));
else
result += applyVisitor(to_string_visitor, value[i]);
/// It is tempting to output DateTime as YYYYMMDDhhmmss, but that would make partition ID
/// timezone-dependent.
}
return result;
}
SipHash hash;
FieldVisitorHash hashing_visitor(hash);
for (const Field & field : value)
applyVisitor(hashing_visitor, field);
char hash_data[16];
hash.get128(hash_data);
result.resize(32);
for (size_t i = 0; i < 16; ++i)
writeHexByteLowercase(hash_data[i], &result[2 * i]);
return result;
}
bool predict_decayed(const predict_orbital_elements_t *orbital_elements, predict_julian_date_t time)
{
double satepoch;
satepoch=DayNum(1,0,orbital_elements->epoch_year)+orbital_elements->epoch_day;
bool has_decayed = false;
if (satepoch + ((16.666666 - orbital_elements->mean_motion)/(10.0*fabs(orbital_elements->derivative_mean_motion))) < time)
{
has_decayed = true;
}
return has_decayed;
}
MergeTreeData::MutableDataPartPtr MergeTreeDataWriter::writeTempPart(BlockWithPartition & block_with_partition)
{
Block & block = block_with_partition.block;
static const String TMP_PREFIX = "tmp_insert_";
/// This will generate unique name in scope of current server process.
Int64 temp_index = data.insert_increment.get();
MergeTreeDataPart::MinMaxIndex minmax_idx;
minmax_idx.update(block, data.minmax_idx_columns);
MergeTreePartition partition(std::move(block_with_partition.partition));
MergeTreePartInfo new_part_info(partition.getID(data), temp_index, temp_index, 0);
String part_name;
if (data.format_version < MERGE_TREE_DATA_MIN_FORMAT_VERSION_WITH_CUSTOM_PARTITIONING)
{
DayNum min_date(minmax_idx.parallelogram[data.minmax_idx_date_column_pos].left.get<UInt64>());
DayNum max_date(minmax_idx.parallelogram[data.minmax_idx_date_column_pos].right.get<UInt64>());
const auto & date_lut = DateLUT::instance();
DayNum min_month = date_lut.toFirstDayNumOfMonth(DayNum(min_date));
DayNum max_month = date_lut.toFirstDayNumOfMonth(DayNum(max_date));
if (min_month != max_month)
throw Exception("Logical error: part spans more than one month.", ErrorCodes::LOGICAL_ERROR);
part_name = new_part_info.getPartNameV0(min_date, max_date);
}
else
part_name = new_part_info.getPartName();
MergeTreeData::MutableDataPartPtr new_data_part = std::make_shared<MergeTreeData::DataPart>(data, part_name, new_part_info);
new_data_part->partition = std::move(partition);
new_data_part->minmax_idx = std::move(minmax_idx);
new_data_part->relative_path = TMP_PREFIX + part_name;
new_data_part->is_temp = true;
/// The name could be non-unique in case of stale files from previous runs.
String full_path = new_data_part->getFullPath();
Poco::File dir(full_path);
if (dir.exists())
{
LOG_WARNING(log, "Removing old temporary directory " + full_path);
dir.remove(true);
}
dir.createDirectories();
/// If we need to calculate some columns to sort.
if (data.hasSortingKey())
data.sorting_key_expr->execute(block);
Names sort_columns = data.sorting_key_columns;
SortDescription sort_description;
size_t sort_columns_size = sort_columns.size();
sort_description.reserve(sort_columns_size);
for (size_t i = 0; i < sort_columns_size; ++i)
sort_description.emplace_back(block.getPositionByName(sort_columns[i]), 1, 1);
ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterBlocks);
/// Sort.
IColumn::Permutation * perm_ptr = nullptr;
IColumn::Permutation perm;
if (!sort_description.empty())
{
if (!isAlreadySorted(block, sort_description))
{
stableGetPermutation(block, sort_description, perm);
perm_ptr = &perm;
}
else
ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterBlocksAlreadySorted);
}
/// This effectively chooses minimal compression method:
/// either default lz4 or compression method with zero thresholds on absolute and relative part size.
auto compression_settings = data.context.chooseCompressionSettings(0, 0);
NamesAndTypesList columns = data.getColumns().getAllPhysical().filter(block.getNames());
MergedBlockOutputStream out(data, new_data_part->getFullPath(), columns, compression_settings);
out.writePrefix();
out.writeWithPermutation(block, perm_ptr);
out.writeSuffixAndFinalizePart(new_data_part);
ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterRows, block.rows());
ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterUncompressedBytes, block.bytes());
ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterCompressedBytes, new_data_part->bytes_on_disk);
return new_data_part;
}
